Skip to content

Agenta

v0.103.0 Feature

This release adds 2 notable features for engineering teams evaluating rollout.

Published 1mo LLM Frameworks
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agent-builder agent-observability agent-orchestration agent-workspace agentic-ai agentic-workflow
+12 more
agents ai-agent ai-agents ai-automation ai-workflow-builder background-agents harness human-in-the-loop llm-evaluation llm-observability mcp self-hosted

Affected surfaces

auth

ReleasePort's take

Moderate signal
editorial:auto 1mo

The release fixes preview recreation in railway links and splits PydanticAI agent spans at the latest user turn. It also removes console logs that leaked tokens, sessions, and PII from the frontend.

Why it matters: Critical security fix: removal of console logs leaking sensitive data (severity 90).

Summary

AI summary

Fixes preview project recreation in railway, splits PydanticAI agent spans in api, and removes token‑leaking console logs from frontend.

Changes in this release

Security Critical

Remove console logs that leak tokens, sessions, and PII.

Remove console logs that leak tokens, sessions, and PII.

Source: llm_adapter@2026-06-10

Confidence: high

 —
Feature Low

Adds safeguards to the table's click handlers.

Adds safeguards to the table's click handlers.

Source: llm_adapter@2026-06-10

Confidence: high

 —
Feature Low

Unify eval loops across the system.

Unify eval loops across the system.

Source: llm_adapter@2026-06-10

Confidence: high

 —
Bugfix Medium

Re-create preview project when a stale railway link fails.

Re-create preview project when a stale railway link fails.

Source: llm_adapter@2026-06-10

Confidence: high

 —
Bugfix Medium

Split PydanticAI agent span messages at the latest user turn.

Split PydanticAI agent span messages at the latest user turn.

Source: llm_adapter@2026-06-10

Confidence: high

 —
Bugfix Medium

Render numeric/discrete evaluator metrics as values, not raw stats.

Render numeric/discrete evaluator metrics as values, not raw stats.

Source: llm_adapter@2026-06-10

Confidence: high

 —
Bugfix Medium

Render LangChain tool calls in the observability drawer.

Render LangChain tool calls in the observability drawer.

Source: llm_adapter@2026-06-10

Confidence: high

 —
Bugfix Medium

Fix playground draft status render for variables.

Fix playground draft status render for variables.

Source: llm_adapter@2026-06-10

Confidence: high

 —
Bugfix Medium

Render playground evaluation workbench changelog entry in docs.

Render playground evaluation workbench changelog entry in docs.

Source: llm_adapter@2026-06-10

Confidence: low

 —
Refactor Low

Move tracing away from legacy endpoints.

Move tracing away from legacy endpoints.

Source: llm_adapter@2026-06-10

Confidence: high

 —
Full changelog

What's Changed

  • [release] v0.102.0 by @github-actions[bot] in https://github.com/Agenta-AI/agenta/pull/4547
  • [feat] Unify eval loops by @junaway in https://github.com/Agenta-AI/agenta/pull/4341
  • fix(railway): re-create preview project when a stale link fails by @mmabrouk in https://github.com/Agenta-AI/agenta/pull/4589
  • fix(api): split PydanticAI agent span messages at the latest user turn by @mmabrouk in https://github.com/Agenta-AI/agenta/pull/4588
  • [fix] Render numeric/discrete evaluator metrics as values, not raw stats by @mmabrouk in https://github.com/Agenta-AI/agenta/pull/4586
  • fix(frontend): remove console logs that leak tokens, sessions, and PII by @ashrafchowdury in https://github.com/Agenta-AI/agenta/pull/4583
  • [fix] Render LangChain tool calls in the observability drawer by @mmabrouk in https://github.com/Agenta-AI/agenta/pull/4582
  • [FE / chore] Move tracing away from legacy endpoints by @ardaerzin in https://github.com/Agenta-AI/agenta/pull/4573
  • [Fix]: playground draft status render for variables by @ashrafchowdury in https://github.com/Agenta-AI/agenta/pull/4585
  • docs(docs): add Playground evaluation workbench changelog entry by @mmabrouk in https://github.com/Agenta-AI/agenta/pull/4596
  • Fix: Adds safeguards to the table's click handlers by @ashrafchowdury in https://github.com/Agenta-AI/agenta/pull/4540

Full Changelog: https://github.com/Agenta-AI/agenta/compare/v0.102.0...v0.103.0

View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track Agenta

Get notified when new releases ship.

Sign up free

About Agenta

LLMOps platform for prompt management, LLM evaluation, and observability. Build, evaluate, and monitor production-grade LLM applications with collaborative prompt engineering.

All releases →

Related context

Earlier breaking changes

  • v0.101.1 Renames `beautified` to `pretty`.
  • v0.100.7 Renames providers table Status column to API Key.

Beta — feedback welcome: [email protected]